Qwen3.5-122b - regolo.ai

Qwen3.5-122B-A10B is a multimodal Mixture-of-Experts model built for native text, image, and video understanding, with 122B total parameters and 10B active parameters per token. The Hugging Face repository provides post-trained weights and configuration files compatible with Hugging Face Transformers, vLLM, SGLang, and KTransformers, which makes the model practical for both research and production deployment. It is positioned for high-utility workloads across reasoning, coding, agent behavior, and visual understanding.

This MoE design delivers near-frontier capabilities in reasoning, coding, and multimodal tasks while optimizing compute.

Example Usage

Qwen3.5-122B-A10B integrates seamlessly with Hugging Face Transformers, vLLM, and similar frameworks.

pip install requestsCode language: Bash (bash)

import requests


api_url = "https://api.regolo.ai/v1/chat/completions"
headers = {
    "Content-Type": "application/json",
    "Authorization": "Bearer YOUR_REGOLO_KEY"
}
data = {
  "model": "qwen3.5-122b",
  "messages": [
    {
      "role": "user",
      "content": "If a train travels 60 km/h for 2 hours and then 80 km/h for 1.5 hours, what is the total distance covered?"
    }
  ],
  "reasoning_effort": "medium"
}

response = requests.post(api_url, headers=headers, json=data)
print(response.json())
Code language: Python (python)

Output

{
  "id": "chatcmpl-8458b75f1bedcde1",
  "created": 1772819136,
  "model": "hosted_vllm/qwen3.5-122b",
  "object": "chat.completion",
  "choices": [
    {
      "finish_reason": "stop",
      "index": 0,
      "message": {
        "content": "To find the total distance covered, we calculate the distance for each part of the trip and then add them together.\n\n**Formula:** $\\text{Distance} = \\text{Speed} \\times \\text{Time}$\n\n1.  **First part of the trip:**\n    *   Speed: $60\\text{ km/h}$\n    *   Time: $2\\text{ hours}$\n    *   Distance: $60 \\times 2 = 120\\text{ km}$\n\n2.  **Second part of the trip:**\n    *   Speed: $80\\text{ km/h}$\n    *   Time: $1.5\\text{ hours}$\n    *   Distance: $80 \\times 1.5 = 120\\text{ km}$\n\n3.  **Total distance:**\n    *   $120\\text{ km} + 120\\text{ km} = 240\\text{ km}$\n\nThe total distance covered is **240 km**.",
        "role": "assistant",
        "reasoning_content": "Thinking Process:\n\n1.  **Analyze the Request:** The user is asking for the total distance covered by a train given two segments of travel with different speeds and durations.\n\n2.  **Identify the Given Information:**\n    *   Segment 1: Speed ($v_1$) = 60 km/h, Time ($t_1$) = 2 hours.\n    *   Segment 2: Speed ($v_2$) = 80 km/h, Time ($t_2$) = 1.5 hours.\n\n3.  **Identify the Formula:** The fundamental formula for distance is $Distance = Speed \\times Time$.\n\n4.  **Calculate Distance for Segment 1:**\n    *   $d_1 = v_1 \\times t_1$\n    *   $d_1 = 60 \\text{ km/h} \\times 2 \\text{ h}$\n    *   $d_1 = 120 \\text{ km}$\n\n5.  **Calculate Distance for Segment 2:**\n    *   $d_2 = v_2 \\times t_2$\n    *   $d_2 = 80 \\text{ km/h} \\times 1.5 \\text{ h}$\n    *   Calculation: $80 \\times 1 = 80$, $80 \\times 0.5 = 40$. $80 + 40 = 120$.\n    *   $d_2 = 120 \\text{ km}$\n\n6.  **Calculate Total Distance:**\n    *   $Total Distance = d_1 + d_2$\n    *   $Total Distance = 120 \\text{ km} + 120 \\text{ km}$\n    *   $Total Distance = 240 \\text{ km}$\n\n7.  **Final Review:** Does the answer make sense?\n    *   First part: 60 km/h for 2 hours -> 120 km. Correct.\n    *   Second part: 80 km/h for 1.5 hours -> 120 km. Correct.\n    *   Total: 120 + 120 = 240 km. Correct.\n\n8.  **Construct the Output:** Present the steps clearly to the user.\n    *   State the formula.\n    *   Calculate the first part.\n    *   Calculate the second part.\n    *   Sum them up.\n    *   State the final answer.cw\n"
      },
      "provider_specific_fields": {
        "stop_reason": null,
        "token_ids": null
      }
    }
  ],
  "usage": {
    "completion_tokens": 788,
    "prompt_tokens": 44,
    "total_tokens": 832
  }
}Code language: JSON / JSON with Comments (json)

Applications & use cases

Multimodal assistants that can understand text, images, and video in a single workflow.
Document understanding and media analysis for reports, visual assets, and mixed-content knowledge bases.
Customer support automation with multilingual coverage across a broad set of languages.
Coding copilots and software engineering assistants for development and debugging tasks.
Agent-style systems that use reasoning and tool-oriented workflows to complete multi-step tasks.
Enterprise knowledge assistants for long-context search, synthesis, and internal documentation workflows.
Research and analysis tasks that benefit from a native 262K-token context and extended long-context configurations.